Free resource

Claude Code commands: the cheat sheet.

Every built-in command, skill, and workflow — what it does and when to reach for it.

Download the PDF → Free to read below. Source: official docs.

Compiled 2026-07-03 against v2.1.199+. Claude Code moves fast — check the official docs for anything added since.

Session Management

/clear	Start fresh conversation, keep project memoryUse when switching to a new task
/compact	Summarize conversation to free contextUse when context gets long mid-task
/resume	Return to a previous conversationPick up where you left off
/branch	Fork conversation to try a different directionExperiment without losing your thread
/rewind	Roll back conversation and code to a checkpointUndo a wrong direction
/context	Visualize what's filling your context windowDiagnose slow/expensive sessions
/btw	Ask a side question without adding to historyQuick lookups that don't belong in the task
/rename	Name the current sessionEasier to find in /resume later
/export	Export conversation as plain text or to clipboardSave a session's output for reuse
/copy	Copy last response or a specific code blockGrab generated code without scrolling

/model	Switch AI model (Opus, Sonnet, Haiku)Use Sonnet for most work; Opus for complex reasoning
/effort	Set reasoning depth: low → medium → high → xhigh → maxHigher effort = slower but better on hard problems
/fast	Toggle fast mode (Opus with faster output)Speed up long tasks without switching models
/advisor	Enable a second model to consult at key momentsBuilt-in second opinion on critical decisions
/usage	Show session cost, token use, and plan limitsCheck spend before it surprises you. Alias: /cost
/plan	Enter plan mode before a large changeClaude proposes, you approve before any code runs
/goal	Set a condition — Claude works until it's metLong-running autonomous tasks

/code-review SKILL skill	Review current diff for bugs and cleanupsRun before every commit. Add --fix to auto-apply
/code-review ultra SKILL skill	Deep multi-agent cloud reviewBefore major PRs or launches
/review	Review a GitHub PR by numberRead-only review of open PRs
/security-review	Scan changes for security vulnerabilitiesAuth, injection, data exposure risks
/simplify SKILL skill	Clean up code — reuse, simplification, efficiencyAfter a feature works; polish before shipping
/diff	Interactive diff viewer — uncommitted changes + per-turnSee exactly what changed before committing

/background	Detach session to run as background agentFree your terminal while Claude keeps working
/fork	Spawn subagent with full conversation contextDelegate a side task while you keep going
/batch SKILL skill	Decompose large codebase changes into parallel unitsMigrate, refactor, or rename across an entire repo
/tasks	View and manage everything running in backgroundMonitor parallel agents and background sessions
/workflows	Watch, pause, or resume running workflowsTrack multi-agent workflow progress
/deep-research WORKFLOW workflow	Fan-out web research, cross-check, synthesize reportMarket research, competitor analysis, technical deep dives

/init	Generate a starter CLAUDE.md for your projectRun first in any new repo
/memory	Edit CLAUDE.md memory files and auto-memoryTune what Claude remembers across sessions
/mcp	Manage MCP server connectionsConnect Gmail, Notion, Stripe, and other services
/permissions	Set allow/ask/deny rules for toolsControl what Claude can do without asking
/hooks	View hook configurations for tool eventsSee what runs before/after Claude's actions
/install-github-app	Install Claude GitHub App + configure ActionsEnable Claude to review PRs automatically
/install-slack-app	Install Claude Slack integrationUse Claude from Slack channels
/add-dir	Grant Claude file access to another directoryWorking across multiple repos in one session

/skills	List all available skillsSee every custom command in your project
/plugin	Install, enable, or disable pluginsAdd new capabilities from the plugin registry
/reload-skills	Reload skills added during the sessionNo restart needed after creating a new skill
/reload-plugins	Reload plugins without restartingApply plugin changes immediately
/loop SKILL skill	Run a prompt on a schedule while session is openMonitor, check-in, or repeat a task at intervals
/schedule	Create cloud-hosted routines that run on a scheduleBriefings, reports, or checks that run without you
/autofix-pr	Cloud agent watches PR and pushes fixes automaticallyCI failures and reviewer comments fixed without you

/run SKILL skill	Launch your app to see a change working liveVerify features in the real app, not just tests
/verify SKILL skill	Confirm a change does what it should, end-to-endBuild + run + observe result automatically
/run-skill-generator SKILL skill	Teach /run and /verify how to build your projectOne-time setup so future runs work automatically
/ultraplan	Draft a plan in browser, then execute remotelyLarge cross-codebase changes with a plan gate
/sandbox	Toggle sandbox mode (no real-world side effects)Safe experimentation in isolated environment

/login	Sign in to your Anthropic account
/logout	Sign out
/status	Show version, model, account, and connectivityQuick health check
/config	Adjust theme, model, output style, and preferences/config theme=dark · /config model=sonnet
/doctor	Diagnose installation and settings issuesPress f to have Claude fix reported issues
/release-notes	View changelog and version historySee what's new after an update
/feedback	Submit feedback or report a bug with session contextAliases: /bug, /share
/team-onboarding	Generate onboarding guide from your session historyGet a teammate set up fast

Every week

One AI workflow, one real use case, one thing you can run today. Free.